Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
dappros avatar

Ethora Mcp Server

  • 6 repo stars
  • Updated July 6, 2026
  • dappros/ethora-mcp-server

Ethora MCP server is an MCP server that exposes Ethora chat, messaging, RAG, and automation APIs to AI coding agents.

About

The Ethora MCP server connects AI coding agents to the Ethora chat and messaging platform through Model Context Protocol, so developers shipping community apps, support bots, or agent-in-the-loop workflows can drive conversations, automation, and retrieval from the editor. It runs as an npm stdio server and expects you to point it at your Ethora API host and supply either app JWT credentials for user bootstrap flows or a B2B server token for tenant-actor authentication. That makes it a strong fit when your product already runs on Ethora and you want the agent to create threads, run RAG-backed replies, or orchestrate messaging without maintaining a custom SDK wrapper. It is less suited as a generic Slack or Discord replacement unless you are committed to Ethora as your messaging backbone. Install the package, set env vars or configure at runtime, and register the server in your agent’s MCP config.

  • stdio MCP package @ethora/mcp-server (v26.5.3) for Claude Code, Cursor, and other MCP clients
  • Configure ETHORA_API_URL or ETHORA_BASE_URL plus optional ETHORA_APP_JWT and ETHORA_B2B_TOKEN for user and tenant auth
  • Runtime configuration via ethora-configure when secrets are not in environment variables
  • Chat ops, AI agents, RAG, and messaging automation against the Ethora platform API
  • Defaults to https://api.chat.ethora.com/v1 when base URL env vars are unset

Ethora Mcp Server by the numbers

  • Data as of Jul 7, 2026 (Skillselion catalog sync)
terminal
claude mcp add --env ETHORA_API_URL=YOUR_ETHORA_API_URL --env ETHORA_BASE_URL=YOUR_ETHORA_BASE_URL --env ETHORA_APP_JWT=YOUR_ETHORA_APP_JWT --env ETHORA_B2B_TOKEN=YOUR_ETHORA_B2B_TOKEN --env ETHORA_MCP_ENABLE_DANGEROUS_TOOLS=YOUR_ETHORA_MCP_ENABLE_DANGEROUS_TOOLS ethora-mcp-server -- npx -y @ethora/mcp-server

Add your badge

Show developers this MCP server is listed on Skillselion. Paste this into your README.

Listed on Skillselion
repo stars6
Package@ethora/mcp-server
TransportSTDIO
AuthRequired
Last updatedJuly 6, 2026
Repositorydappros/ethora-mcp-server

What it does

Wire your coding agent into Ethora chat, messaging automation, RAG, and in-app AI agents without hand-rolling REST calls.

Who is it for?

Best when you're already on Ethora and want agent-driven chat workflows, RAG, or B2B tenant automation during feature development.

Skip if: Skip if you only need a public website chat widget with no Ethora tenant, or developers and want a single skill file instead of a live API MCP bridge.

What you get

After registration, your agent can configure Ethora credentials and perform chat ops and automation through typed MCP tools instead of ad-hoc HTTP scripts.

  • Registered stdio MCP server the agent can invoke for Ethora chat and automation tools
  • Documented env-based or runtime ethora-configure credential setup
  • Agent-callable Ethora API operations without custom HTTP glue in every task

By the numbers

  • Package version 26.5.3 on npm as @ethora/mcp-server
  • Transport: stdio
  • Default API base resolves to https://api.chat.ethora.com/v1 when env URLs are omitted
README.md

Ethora MCP Server (Model Context Protocol)

npm Node License

Add to Cursor Install in VS Code Install in VS Code Insiders

One-click install for Cursor and VS Code (buttons above). For Claude Code, Claude Desktop, GitHub Copilot, Gemini CLI, Codex CLI, Windsurf, and Cline, see Using with MCP Clients below.

An MCP (Model Context Protocol) CLI/server that connects popular MCP clients to the Ethora platform — an open-source chat & messaging platform with a built-in AI agent / chatbot framework. This runs locally on a developer machine via stdio rather than as a hosted Ethora service.
Use it from Cursor, VS Code MCP, Claude Desktop, or Windsurf/Cline to manage apps and chat rooms, broadcast messages, deploy AI agents / chatbots with RAG sources, and automate B2B provisioning workflows. (ERC-20 wallet tools are also included — see the tool list below.)

Part of the Ethora SDK ecosystem — see all SDKs, tools, and sample apps. Follow cross-SDK updates in the Release Notes.


✨ What you get

  • Talk to the Ethora platform directly from your IDE or AI agent client (Cursor, VS Code MCP, Claude Desktop, Windsurf / Cline).
  • Both user-auth flows (login/register, files, owner/admin endpoints) and B2B / app-token flows (tenant provisioning, broadcast jobs, async user batches, AI bot config).
  • Built-in recipes, prompts, and generators for the most common Ethora workflows (Vite/Next chat-component setup, B2B bootstrap, AI-bot enablement, RAG sources).
  • Standard tool response envelope ({ ok, ts, meta, data | error }) so agent code can reason about success/failure consistently.

🚦 Just trying it? (60-second quickstart)

Don't read the auth modes yet. Once the server is connected in your client, ask your agent to run, in order:

  1. ethora-doctor — confirms the server is up and can reach the Ethora API. No credentials needed.
  2. ethora-configure with your appJwtethora-auth-use-userethora-user-login with an email + password.
  3. ethora-app-list — you're in; this lists your apps.

That's the local-developer path. Need server-side automation instead? Jump to B2B mode. Lost at any point, call ethora-help — it reads your current state and tells you the next call.

🔐 Two typical usage modes

1) User Auth mode

Best for:

  • developers trying Ethora locally
  • tenant admins / app owners using MCP manually
  • flows that start with ethora-user-login

How it works:

  • configure ETHORA_APP_JWT once for login/register bootstrap
  • switch to ethora-auth-use-user
  • call ethora-user-login
  • use user-auth tools such as files and legacy owner/admin endpoints

2) B2B mode

Best for:

  • permanent backend integrations
  • partner provisioning flows
  • autonomous agents operating Ethora without a human user session

How it works:

  • configure ETHORA_B2B_TOKEN
  • switch to ethora-auth-use-b2b for explicit tenant-actor /v2/apps/:appId/... routes
  • optionally switch into ethora-auth-use-app after ethora-app-select when you want app-scoped convenience routes powered by appToken

Rule of thumb:

  • first-time local use usually starts with User Auth
  • repeatable automation usually starts with B2B, then often moves into app-token mode for one selected app

Prompts & Resources (P2: dev-facing docs)

  • Resources (loadable docs into context)
    • ethora://docs/auth-map — appJwt vs appToken vs b2bToken
    • ethora://docs/chat-component/quickstart — Vite/Next quickstart + replacing demo tokens
    • ethora://docs/sdk-backend/quickstart — backend integration quickstart
    • ethora://docs/recipes — common tool sequences (broadcast/sources/files/bot)
  • Prompts
    • ethora-auth-map
    • ethora-vite-quickstart
    • ethora-nextjs-quickstart
    • ethora-backend-sdk-quickstart
    • ethora-recipes

Generators (no shell, no file writes)

  • ethora-generate-chat-component-app-tsx — ready-to-paste App.tsx snippet for @ethora/chat-component
  • ethora-generate-env-examples.env.example templates for:
    • frontend chat component
    • backend SDK integration
    • MCP usage (ETHORA_API_URL, ETHORA_APP_JWT, ETHORA_B2B_TOKEN)
  • ethora-generate-b2b-bootstrap-runbook — minimal “call these MCP tools in order” runbook for B2B bootstrap

Tip: to list runnable recipes without calling ethora-help, call ethora-run-recipe with goal: "auto" and omit recipeId.

  • Session / Config

    • ethora-configure — set API URL plus App JWT / B2B token / appToken for this MCP session
    • ethora-status — show configured API URL, active auth mode, and which credentials are present
    • ethora-help — task-oriented help (recommended next calls + “one-click recipes” based on current state)
    • ethora-run-recipe — execute a built-in recipe by id (sequential steps; no shell, no file writes)
    • ethora-doctor — validate config + ping the configured Ethora API for both user and B2B usage
    • ethora-app-select — select current appId and optionally set appToken
    • ethora-auth-use-app — switch to app-token auth mode for app-scoped operations
    • ethora-auth-use-user — switch to user-session auth mode
    • ethora-auth-use-b2b — switch to tenant-actor B2B x-custom-token auth mode
  • Chats (v2)

    • ethora-chats-broadcast-v2 — enqueue broadcast job using app-token auth or B2B + explicit appId
    • ethora-chats-broadcast-job-v2 — get broadcast job status/results using app-token auth or B2B + explicit appId
    • ethora-wait-broadcast-job-v2 — poll broadcast job until completed/failed using app-token auth or B2B + explicit appId
    • ethora-chats-message-v2 — send a test/automation message through the app chat surface (requires app-token auth)
    • ethora-chats-history-v2 — read persisted automation/test history for private or group sessions (requires app-token auth)
  • Users (v2 async batch)

    • ethora-users-batch-create-v2 — create async users batch job (requires B2B auth)
    • ethora-users-batch-job-v2 — get users batch job status/results (requires B2B auth)
    • ethora-wait-users-batch-job-v2 — poll users batch job until completed/failed (requires B2B auth)
  • Files (v2)

  • Bot / Agent (v2)

    • ethora-bot-get-v2 — get bot status/settings using app-token auth or B2B + explicit appId

    • ethora-bot-update-v2 — update bot settings using app-token auth or B2B + explicit appId

    • ethora-bot-enable-v2 — enable bot using app-token auth or B2B + explicit appId

    • ethora-bot-disable-v2 — disable bot using app-token auth or B2B + explicit appId

    • ethora-bot-widget-v2 — get widget/embed config and public widget URL metadata (app-token auth)

    • ethora-agents-list-v2 — list reusable saved agents for the current app owner (app-token auth)

    • ethora-agents-get-v2 — get one reusable saved agent (app-token auth)

    • ethora-agents-create-v2 — create a reusable saved agent (app-token auth)

    • ethora-agents-update-v2 — update a reusable saved agent (app-token auth)

    • ethora-agents-clone-v2 — clone a reusable saved agent (app-token auth)

    • ethora-agents-activate-v2 — bind a saved agent as the active bot for the selected app (app-token auth)

    • ethora-bot-message-v2 — compatibility alias for ethora-chats-message-v2

    • ethora-bot-history-v2 — compatibility alias for ethora-chats-history-v2

    • ethora-files-upload-v2 — upload files (requires user auth)

    • ethora-files-get-v2 — list/get files (requires user auth)

    • ethora-files-delete-v2 — delete file by id (requires user auth)

  • Sources

    • ethora-sources-site-crawl — crawl a URL (requires user auth)
    • ethora-sources-site-reindex — reindex URL by urlId (requires user auth)
    • ethora-sources-site-delete-url — delete by URL (requires user auth)
    • ethora-sources-site-delete-url-v2 — batch delete URLs (requires user auth)
    • ethora-sources-docs-upload — upload docs for ingestion (requires user auth)
    • ethora-sources-docs-delete — delete ingested doc by id (requires user auth)
    • ethora-sources-site-crawl-v2 — crawl a URL using app-token auth or B2B + explicit appId
    • ethora-sources-site-reindex-v2 — reindex URL by urlId using app-token auth or B2B + explicit appId
    • ethora-sources-site-crawl-v2-wait — single-call long-timeout helper for crawl (app-token auth)
    • ethora-sources-site-reindex-v2-wait — single-call long-timeout helper for reindex (app-token auth)
    • ethora-sources-site-list-v2 — list crawled site sources and current tags using app-token auth or B2B + explicit appId
    • ethora-sources-site-tags-update-v2 — set/update tags for a crawled site source using app-token auth or B2B + explicit appId
    • ethora-sources-site-delete-url-v2 — delete one crawled URL by URL using app-token auth or B2B + explicit appId
    • ethora-sources-site-delete-url-v2-batch — batch delete crawled source records by id using app-token auth or B2B + explicit appId
    • ethora-sources-docs-upload-v2 — upload docs for ingestion using app-token auth or B2B + explicit appId
    • ethora-sources-docs-list-v2 — list indexed documents and current tags using app-token auth or B2B + explicit appId
    • ethora-sources-docs-tags-update-v2 — set/update tags for an indexed document using app-token auth or B2B + explicit appId
    • ethora-sources-docs-delete-v2 — delete doc by id using app-token auth or B2B + explicit appId
  • Auth & Accounts

    • ethora-user-login — login user (email + password)
    • ethora-user-register — register user (email + first/last name)
  • Applications

    • ethora-app-create — create app
    • ethora-app-update — update app
    • ethora-app-delete — delete app
    • ethora-app-list — list apps
    • ethora-b2b-app-create — create app using B2B auth (x-custom-token)
    • ethora-b2b-app-bootstrap-ai — create app → index sources → configure/enable bot, including runtime LLM selection (B2B automation)
    • ethora-app-tokens-list-v2 — list app token metadata (B2B auth)
    • ethora-app-tokens-create-v2 — create new app token (returned once) (B2B auth)
    • ethora-app-tokens-rotate-v2 — rotate token (revoke old, return new once) (B2B auth)
    • ethora-app-tokens-revoke-v2 — revoke token by tokenId (idempotent) (B2B auth)
    • ethora-b2b-app-provision — create app + create tokens + provision rooms + configure bot, including runtime LLM selection (B2B orchestrator)
  • Chat & Rooms

    • ethora-app-get-default-rooms — list default rooms
    • ethora-app-get-default-rooms-with-app-id — rooms for a given app
    • ethora-app-create-chat — create chat for app
    • ethora-app-delete-chat — delete chat
  • Wallet

    • ethora-wallet-get-balance — get balance
    • ethora-wallet-erc20-transfer — send ERC-20 tokens

Tool names above reflect the functional areas exposed by the server. Your exact tool names may vary slightly by version; run the client’s “list tools” to confirm.

settings login

📦 Install / Run

Pre-requisites

Before you begin, ensure you have the following:

  • Node.js installed on your system (recommended version 18.x or higher).

Install

The server is distributed as an npm package and is typically launched by MCP clients via npx:

npx -y @ethora/mcp-server

No global install is required.


🔐 Configuration (env vars)

This MCP server supports both the local user-auth flow and the server-side B2B flow.

Core values:

  • Ethora API URL (where to send requests)
  • Ethora App JWT (used only for login/register bootstrap in user-auth mode)
  • Ethora B2B Token (used for tenant-actor server-to-server flows)

You can provide these either:

  • via env vars, or
  • at runtime via the ethora-configure tool (in-memory; resets when MCP process restarts)

Supported env vars

  • ETHORA_API_URL: full API URL (example: https://api.chat.ethora.com/v1, http://localhost:8080/v1)
  • ETHORA_BASE_URL: base host URL (example: https://api.chat.ethora.com, http://localhost:8080)
    If provided, the server will default to .../v1.
  • ETHORA_APP_JWT: App JWT string, usually starting with JWT ...
  • ETHORA_B2B_TOKEN: B2B server token for x-custom-token auth (JWT with type=server)
  • ETHORA_MCP_ENABLE_DANGEROUS_TOOLS: enable destructive tools (default: disabled). Set to true to expose:
    • app deletion tools
    • wallet transfer tools
    • bulk delete tools

Security: never commit App JWTs, B2B tokens, or appTokens to git. Configure them via env vars, the MCP client secret store, or your own backend.


🧱 Standard response envelope (tools)

All tools return JSON in a consistent envelope:

  • Success: { ok: true, ts, meta, data }
  • Error: { ok: false, ts, meta, error }, where error includes:
    • code: stable string (prefer API code, otherwise inferred)
    • httpStatus: HTTP status when the failure came from an API call
    • requestId: request/correlation id if returned by API
    • hint: 1-line “what to do next”

🚀 Using with MCP Clients

Every client runs the same thing — npx -y @ethora/mcp-server over stdio. One-click buttons exist for Cursor and VS Code (top of this README). For the rest it's a short config block or a one-line command.

Cursor

Use the Add to Cursor button above, or manually: Settings → MCP → Add new global MCP server:

{
  "mcpServers": {
    "ethora": {
      "command": "npx",
      "args": ["-y", "@ethora/mcp-server"]
    }
  }
}

VS Code (and GitHub Copilot)

Use the Install in VS Code button above, or add a .vscode/mcp.json file (project-level) — note the key is servers:

{
  "servers": {
    "ethora": {
      "command": "npx",
      "args": ["-y", "@ethora/mcp-server"]
    }
  }
}

GitHub Copilot's agent mode in VS Code reads this same .vscode/mcp.json — no separate setup. (For a user-level install instead, put the servers block under "mcp" in your User Settings JSON.)

Claude Code

One command:

claude mcp add ethora -- npx -y @ethora/mcp-server

Add --scope user to make it available in every project. Verify with claude mcp list.

To pre-configure credentials, pass them as env vars with -e (recommended over the ethora-configure tool — see note below):

claude mcp add ethora \
  -e ETHORA_API_URL=https://api.chat.ethora.com/v1 \
  -e ETHORA_B2B_TOKEN=<your-b2b-token> \
  -- npx -y @ethora/mcp-server

Secrets note: prefer env vars (above) or your MCP client's secret store for credentials. The ethora-configure tool also works, but it passes secrets as tool arguments, which means they end up in the conversation transcript. Use it for quick local testing, not for tokens you care about.

Claude Desktop

Settings → Developer → Edit Config, open claude_desktop_config.json:

{
  "mcpServers": {
    "ethora": {
      "command": "npx",
      "args": ["-y", "@ethora/mcp-server"]
    }
  }
}

Gemini CLI

Add to ~/.gemini/settings.json (global) or .gemini/settings.json (per project):

{
  "mcpServers": {
    "ethora": {
      "command": "npx",
      "args": ["-y", "@ethora/mcp-server"]
    }
  }
}

Codex CLI

Add to ~/.codex/config.toml — note the table name is mcp_servers (underscore; mcp-servers is silently ignored):

[mcp_servers.ethora]
command = "npx"
args = ["-y", "@ethora/mcp-server"]

Windsurf

Settings → Cascade → MCP Servers → View raw config (~/.codeium/windsurf/mcp_config.json):

{
  "mcpServers": {
    "ethora": {
      "command": "npx",
      "args": ["-y", "@ethora/mcp-server"]
    }
  }
}

Cline

Open the MCP servers panel and edit cline_mcp_settings.json:

{
  "mcpServers": {
    "ethora": {
      "command": "npx",
      "args": ["-y", "@ethora/mcp-server"]
    }
  }
}

🧪 Quick test

After the server shows as connected in your client:

  • Run list tools (client command) to verify Ethora tools are available.
  • Check config/connectivity: call ethora-doctor (or ethora-status)
  • For a first local/manual test:
    • call ethora-configure with apiUrl / appJwt
    • call ethora-auth-use-user
    • call ethora-user-login
    • then try ethora-app-list or ethora-wallet-get-balance
  • For a server-side/B2B test:
    • call ethora-configure with apiUrl / b2bToken
    • call ethora-auth-use-b2b
    • then try ethora-b2b-app-create or ethora-app-tokens-list-v2

🧭 P1: B2B “create app → index sources → deploy bot” in one call

Pre-reqs:

  • Configure ETHORA_API_URL (or call ethora-configure)
  • Configure ETHORA_B2B_TOKEN (or call ethora-configure with b2bToken)
  • Ensure your Ethora backend is configured with AI service URL/secret (for bot activation)

Suggested flow:

  • Call ethora-auth-use-b2b
  • Call ethora-b2b-app-bootstrap-ai with:
    • displayName
    • optional savedAgentId
    • optional crawlUrl
    • optional docs[] (base64)
    • enableBot: true
    • optional llmProvider
    • optional llmModel

It will:

  • create the app (B2B)
  • set current app context (best-effort)
  • index sources via /v2/sources/* (app-token auth)
  • configure and/or enable bot (best-effort)

Example payloads

Minimal (create app only):

{
  "displayName": "Acme AI Demo",
  "setAsCurrent": true
}

Create app + crawl a website + enable bot:

{
  "displayName": "Acme AI Demo",
  "savedAgentId": "6790abc1234567890def1111",
  "crawlUrl": "https://example.com",
  "followLink": true,
  "enableBot": true,
  "botTrigger": "/bot",
  "llmProvider": "openai",
  "llmModel": "gpt-4o-mini"
}

Create app + upload docs + enable bot:

{
  "displayName": "Acme AI Demo",
  "docs": [
    {
      "name": "faq.pdf",
      "mimeType": "application/pdf",
      "base64": "BASE64_PDF_CONTENT_HERE"
    }
  ],
  "enableBot": true,
  "llmProvider": "openai",
  "llmModel": "gpt-4o-mini"
}

Provision app + token + default rooms + bot settings:

{
  "displayName": "Acme Support",
  "savedAgentId": "6790abc1234567890def1111",
  "tokenLabels": ["default", "staging"],
  "rooms": [
    { "title": "General" },
    { "title": "Support", "pinned": true }
  ],
  "enableBot": true,
  "botTrigger": "/bot",
  "botPrompt": "You are the Acme support assistant.",
  "botGreetingMessage": "Hello. How can I help?",
  "llmProvider": "openai",
  "llmModel": "gpt-4o-mini"
}

Provider/model note:

  • Common values are openai and openai-compatible.
  • The effective provider/model must also be enabled by your Ethora backend + AI service environment.

🤖 App automation loop

Once you already have an app selected with appToken auth:

  • call ethora-auth-use-app
  • call ethora-bot-get-v2 to inspect current bot status and prompt settings
  • call ethora-sources-site-list-v2 and ethora-sources-docs-list-v2 to inspect indexed sources
  • call ethora-sources-site-tags-update-v2 or ethora-sources-docs-tags-update-v2 to organize retrieval by tags
  • call ethora-chats-message-v2 / ethora-chats-history-v2 if your backend exposes the chat automation surface on the same API host

Example: apply retrieval tags to a crawled source

{
  "sourceId": "6790abc1234567890def1234",
  "tags": ["support", "faq", "billing"]
}

Example: apply retrieval tags to an indexed document

{
  "docId": "6790abc1234567890def1235",
  "tags": ["support", "faq"]
}

🛡️ Security notes

  • Never hardcode API keys in shared config. Prefer client-side secret stores.
  • Use least privilege keys and consider allowlists/rate limits on your Ethora backend.
  • Rotate credentials regularly in production use.

CI security scans (report-only)

This repo runs report-only scans on pushes/PRs:

  • gitleaks for secret scanning
  • semgrep for basic SAST

🧰 Development

Clone and run locally:

git clone https://github.com/dappros/ethora-mcp-server.git
cd ethora-mcp-server
npm install
npm run build
npm start

Suggested scripts (if not present):

{
  "scripts": {
    "build": "tsc -p .",
    "start": "node dist/index.js",
    "dev": "tsx src/index.ts"
  }
}

❓ Troubleshooting

  • Client can’t connect: Ensure npx @ethora/mcp-server runs locally without errors. Check Node ≥ 18.
  • Auth errors: Verify ETHORA_BASE_URL and any required secrets are set in the client’s environment.
  • Tools missing: Restart the MCP client and inspect server logs for registration errors.
  • Network: Confirm outbound access from the IDE to your Ethora host.

🔗 Related Repos


📜 License

See LICENSE.

Recommended MCP Servers

How it compares

Live Ethora API bridge over MCP, not a static Claude skill for prompt-only chat design.

FAQ

Who is io.github.dappros/ethora-mcp-server for?

It is for and small-team developers building on the Ethora messaging platform who want their Claude Code, Cursor, or Codex agent to run chat and automation against real Ethora APIs.

When should I use io.github.dappros/ethora-mcp-server?

Use it during the build phase when you are wiring agent tooling or backend integrations for Ethora-hosted chat, AI agents, or RAG-backed messaging features.

How do I add io.github.dappros/ethora-mcp-server to my agent?

Install @ethora/mcp-server from npm, set ETHORA_API_URL or ETHORA_BASE_URL and the appropriate JWT or B2B token env vars, then add the stdio MCP server entry in your agent’s MCP configuration.

Communicationagentsautomation

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.